Artificial intelligence (AI) chatbots are less competent in languages other than English, raising concerns of exacerbating existing bias in global commerce and innovation. This issue threatens to limit the potential of polyglot AI helpers like ChatGPT that could bridge language barriers. However, the dominance of English in global commerce is real, and non-English speakers are often punished professionally. While AI researchers are working on testing and rectifying the chatbots’ language skills, the mostly English-speaking chatbots continue to spread. Tests show that these AI systems are good at translating other languages into English, but they struggle to convert English into other languages. Non-English languages, especially those with non-Latin scripts, pose the most significant challenges. Although the technology’s limitations are evident, workers worldwide are relying on chatbots for help in various areas.
ChatGPT is powered by large language models, and studies show that it struggles with fluent mixing of languages in the same utterance. If the technology continues to work well in English, it could force people hoping to earn a place in the global economy to learn the language. This would further increase the imposition and influence of English that began with the British Empire. However, Pascale Fung, director of the Center for AI Research at Hong Kong University of Science and Technology, is part of a group of AI researchers testing the language skills of chatbots and sounding the alarm. According to Fung, AI change is essential to prevent the primacy of English, but this change requires more than just translation tools.
Pascale Fung is a computer scientist and director of the Center for AI Research at the Hong Kong University of Science and Technology.
ChatGPT is an AI chatbot that is less capable in languages other than English, spreading concern about exacerbating existing bias in global commerce and innovation.